community environmental noise and the built environment in...

Page | i

Dalhousie University – Environmental Science

Community environmental noise and the built environment in two Halifax neighbourhoods

Supervised by Daniel Rainham & Marek Roland- Mieszkowski

Gavin King 4/11/2008

Page | ii

Table of Contents Acknowledgments............................................................................................................... 1

1. Abstract ..................................................................................................................... 2

2. Introduction .............................................................................................................. 3

2.1 Objectives and Scope ............................................................................................ 6

3. Literature Review ...................................................................................................... 8

3.1 Sampling Strategies ............................................................................................. 10

3.2 Sampling Periods ................................................................................................. 12

3.3 Health Related Values ......................................................................................... 13

3.4 Data Analysis ....................................................................................................... 16

4. Methods .................................................................................................................. 17

4.1 Study areas .......................................................................................................... 17

4.2 Instrumentation .................................................................................................. 20

4.3 Sampling periods ................................................................................................. 21

4.4 Sample Points ...................................................................................................... 21

4.5 Analysis ................................................................................................................ 22

4.6 Health Comparisons ............................................................................................ 24

5. Data Presentation and Analysis .............................................................................. 25

5.1 Area one .............................................................................................................. 26

5.2 Area two .............................................................................................................. 36

5.3 Analysis ................................................................................................................ 45

6. Limitations, Recommendations and Conclusions ................................................... 49

7. Reference List .......................................................................................................... 52

Page | iii

Table of Figures Figure 1. Map of study areas. .......................................................................................... 19

Figure 2. Study equipment. .............................................................................................. 20

Figure 3. Area 1 LA1 values ............................................................................................... 29

Figure 4. Area 1 LA90 values. ............................................................................................. 29

Figure 5. Area 1 full 24-hour LAeq values with adjustment. .............................................. 31

Figure 6. Histogram plot of Area one data points. .......................................................... 32

Figure 7. Kruskal-Wallis test for area 1 sample sites. ...................................................... 33

Figure 8. Boxplot and 95% confidence interval for site 6, Area 1. .................................. 34

Figure 9. Boxplot and 95% confidence interval for site 6, Area 1. .................................. 35

Figure 10. Area 2 LA1 values. ............................................................................................ 40

Figure 11. Area 2 LA90 values. ........................................................................................... 40

Figure 12. Area 2 full 24 hourLAeq values with adjustment. ............................................. 41

Figure 13. Histogram plot of Area 2 data points.............................................................. 42

Figure 14. Boxplot and 95% confidence interval for site 6 Area 2. ................................. 43

Figure 15. Boxplot and 95% confidence interval for site 3 Area 2. ................................. 44

Figure 16. Area 1 & 2 adjusted LAeq values. ..................................................................... 45

Figure 17. Areas 1 & 2 LA90 and Adjusted LAeq values. ...................................................... 46

Figure. 18 Levene’s test of equality of error variances. ................................................... 48

Page | iv

Table of Tables

Table 1. EPA guidelines for environmental noise............................................................14

Table 2. Italian legislation limits for environmental noise............................................... 15

Table 3. Summary table of various statistical values for Area 1. ..................................... 27

Table 3 (Continued).. ........................................................................................................ 28

Table 4. Selected Area 1 statistical values distributed by sample time period. .............. 30

Table 5. Summary table of various statistical values for Area 2 ...................................... 37

Table 5 (continued) .......................................................................................................... 38

Table 6. Slected Area 2 statistical values distributed by sample time period. ................ 39

Table 7.. LAeq and Adjusted LAeq values con values for human health ............................. 48

Page | 1

ACKNOWLEDGMENTS

I would like to recognize the kind assistance, direction, and support provided by my

supervisors Daniel Rainham and Marek Roland - Mieszkowski during the completion of this

project. Without their direction and expertise this project could not have happened. As well, I

would like to thank the Dalhousie GIS centre for providing excellent advice with ArcGIS,

Musicstop and Jack Julian of CBC Radio for assistance with equipment. As well, thank you to my

friends, family and classmates for mental and moral support and generous assistance

throughout all stages of this project.

Gavin King

Gavin King Dalhousie University

[email protected] April 11 2008

Page | 2

1. ABSTRACT

A study of two Halifax neighbourhoods was conducted to examine the effect of the built

environment and land use on levels of environmental noise during January and February of

2008. The study areas were identified using dissemination areas of the Canadian Census

program, land use information, air photography and ground truthing. The first study area was a

residential area consisting of two- to three-story single family homes and the second was a

mixed-use region with apartment buildings, as well as commercial and institutional

development. Study areas were gridded into six areas and a random sample point in each grid

area was identified using ArcGIS. Each sample point was sampled four times over the 24-hour

day, resulting in a total of 24 samples for each of the two areas, with one sample taken per

hour. Study results demonstrate that the mixed-use area had statistically significant higher

levels of environmental noise then the residential area. Study results were compared to health

guidelines from the EPA and Italian Government and found to exceed the allowable limits.

Page | 3

2. INTRODUCTION

Human and community health are greatly affected by the built environment. The term

“built environment” refers to the portion of the physical environment constructed by humans,

for humans, and it includes buildings, roads, and other transportation systems, as well as open

spaces like parks and sports fields (1). The structure and use of the built environment are

linked to various types of negative health impacts, including cardiovascular and pulmonary

disease, obesity, and asthma, as well behavioural impacts such as depression, stress and

annoyance (1;2). Approximately 80% of the population of North America resides in an urban

environment, and is exposed to various pollutants produced by human activities including

environmental noise, particulate matter, vehicle emissions, and chemical runoff. Therefore the

urban environment has become an important focus of current environmental, human health

and epidemiological research.

Environmental noise, as with other types of anthropogenic area pollution results from

the high population concentrations, correspondingly high concentrations of vehicles, and

intense industrial and commercial development found in the urban environment. Main sources

of environmental noise include road traffic, construction and public works, industrial and

institutional activity, and social and economic activities (3;4). In the urban environment, these

activities are brought into close proximity to the human population increasing exposure and

possible health effects.

Environmental noise has been linked by epidemiological research to health outcomes

including increased levels of hypertension and high blood pressure (5), lowered cognitive ability

Page | 4

(6), and increased occurrence of cardiovascular disease (7). Qualitative studies have shown

that environmental noise is considered to be the most annoying of all types of urban pollution

(3;8), interfering with enjoyment of daily activities and causing loss of sleep and rest. This can

compel people to migrate to other cities or into less populated areas, like suburbs, contributing

to urban sprawl and increasing environmental noise and other pollution.

Much environmental noise research is concentrated primarily on occupational

exposures for the development of regulation. These regulations prescribe the amount of time

workers are allowed to be exposed to specific sound pressure levels (SPL) and outline the type

of noise protection workers are entitled to. Little work has been conducted into the

relationship of environmental noise and the built environment and the resulting environmental

noise.

Environmental noise is considered by the World Health Organization (WHO) to be any

noise which is not occupation related, either indoor or outdoor(9). WHO identifies the main

sources of indoor noise as ventilation systems, home appliances, office machines and

neighbourhood activities. Outdoor sources of environmental noise include social noise,

commercial and domestic noises, music, playgrounds, sporting events, pets and transportation

infrastructure. Difficulties with the definition, measurement and control, as well as insufficient

understanding about the health effects of noise on people, have resulted in an inability to

effectively ameliorate environmental noise at levels potentially harmful to health (9).

As with many Canadian urban centres, Halifax, through the 2007 Regional Municipal

Planning Strategy (RMPS), is planning to intensify urban development by combining land-use

Page | 5

types to create mixed-use regions with a focus development into core areas. There are good

reasons for densification of urban core areas. First, Halifax will require approximately 18 000

hectares of new residential development by 2010 if current projected growth rates continue

(10). Second, continued urban sprawl will result in increased costs for the provision of

municipal services including water and sewage piping, emergency services, as well as road

development and maintenance.

Urban sprawl significantly impacts the physical environment of a city. Studies have

shown urban sprawl to be implicated in soil loss (11), fragmentation and loss of local forest

areas (12), as well as local arable land (13). Urban sprawl also influences human health through

direct and indirect mechanism. For example, increased reliance on automobiles as a form of

transportation has been shown to increase air pollution concentrations resulting in increased

prevalence of cardiovascular disease (14). Automobile reliance has also been shown to reduce

levels of physical activity which has led to increased prevalence of diabetes and levels of obesity

(15-17). These issues make the planning strategy to intensify development in core areas very

important for the health and sustainability of our communities.

Conversely, intensification of land-use can also have unhealthy effects on human

populations and the environment. Several studies have identified areas of health concern that

can result from increased urbanization which include social alienation and psychosocial health

problems, increased risk of disease resulting from living in close quarters with populations of

animal vectors, like rats and fleas, and increased exposure to pollutants like particle pollution

and environmental noise (18-20). It is important to understand these health issues at both the

Page | 6

planning and regulation level because while intensification is necessary for the health of cities

and possibly the world, we need to understand how our actions will affect the human residents

of the cities and properly protect their health and safety.

2.1 Objectives and Scope

The objective of this study is to sample and analyze environmental noise data characteristic

to two distinct urban land uses. By examining levels of environmental noise in two forms of the

urban built environment, the study provides information relevant to urban planners and public

health professionals. The project is timely given that little is known about the potential for

human health impacts arising from environmental noise as a result of current plans to intensify

development of the urban core.

One of the study neighbourhoods is an older-style, traditional residential neighbourhood

with few roads and predominantly free standing single family dwellings. The second study area

is a modern-style neighbourhood which has a high road density, and mixed land-use including

residential apartment buildings, commercial enterprises and first floor commercial

development with residential development above. This study is one of the first to examine

environmental noise in Halifax.

The scope of this project is limited to these two representative areas as it is not feasible

to measure environmental noise across the whole of Halifax Regional Municipality (HRM)

because of time and financial constraints. The project is also limited temporally as it focuses

only on weekdays, and does not include seasonal influences. This study examines the complete

Page | 7

24-hour day in order to fully estimate the variation in the character and volume of

environmental noise throughout the day and night.

Page | 8

3. LITERATURE REVIEW

Environmental noise has a documented impact on the health and well-being of

individuals. Although research has traditionally been focused on occupational exposure (21),

recent studies have examined the effects of environmental noise outside the work

environment, including how it affects humans at the population level. The psychological and

physiological effects of exposure to environmental noise include sleep loss, annoyance, hearing

loss, cardiovascular problems, and depressed task performance (4-6).

Sound is comprised of alternating compressions and expansions of the medium through

which the sound wave travels. The volume of the sound is expressed as a slight positive or

negative deviation of atmospheric pressure, the greater the deviation the louder the sound,

while the frequency of the cycle of alterations determines the pitch (4). The human ear can

normally detect frequencies between 20 hertz (Hz) and 20 kilohertz (kHz); however this range

can be affected by damage resulting from loud sounds or illness. A low frequency sound is

heard as a deep hum, while a high frequency sound is heard as a squeak.

Sound is measured by comparing the logarithm of a given sound to a reference sound

pressure, and is expressed on a logarithmic decibel (dB) scale. When assessing the impact of

sound it is important to note that the difference between 60 dB and 70 dB is a ten-fold increase

in volume and the impact of the sound will also be ten-fold greater (4). The human ear does

not register sound equally on all frequencies, so the effects of sound on humans cannot be

accurately assessed without weighting the measurements to reflect human hearing sensitivity.

Page | 9

A weighting system called A-weighting is used to adjust results in studies examining human

impacts from environmental noise to better reflect human hearing [dB (A)] (4).

Environmental noise research is centred in a few regions across the globe,

reflecting the local societal evaluation of the significance of the dangers and annoyance of

environmental noise. The European Union and Hong Kong are two prominent areas of sound

research, and both regions regulate environmental noise. Italy, for example, has environmental

noise exposure limits set by federal legislation with identified acceptable levels of sound for

given land-uses, for both daytime and night-time (22). Other cities, including San Francisco,

USA (8), Hurghada, Egypt (23), and Valdivia, Chile (24) have also been studied in recent years,

demonstrating the developing awareness of environmental noise and its impact.

A review of available documentation shows that Canadian interests lie primarily in the

regulation of occupational noise and noise emitted from consumer goods. Powered equipment

and vehicles must conform to federal regulations for levels of acceptable sound emission. For

example, Schedule v.1 (Section 5) Noise Emissions - Standard 1106 of the Motor Vehicle Safety

Regulations provides allowable interior and exterior sound levels for different classes of

vehicles. However these regulations do not apply to “after sale” conditions, such as modified

mufflers. Federal regulations also control noise from inter-provincial transportation systems

including trains, highways, aircraft, and waterways. As well there are provisions for national

occupational health and safety guidelines found in the Canada Labour Code, Part II, (R.S.C.

1985, c. L-2) Canada Occupational Safety and Health Regulations, (SOR/86-304) Section

7.4(1)(b).

Page | 10

Provincial governments regulate the operational noise levels of equipment, vehicles

and other commercial goods, and produce environmental noise guidelines for land-use types

and for provincial roads, as well as setting occupational health and safety standards under the

Occupational Health and Safety Act (S.N.S. 1996, c.7). Municipal governments are responsible

for environmental noise regulation based on disturbance, with bylaws controlling noise-

emitting events. Examples include the Halifax Regional Municipality (HRM) By-Law N-200 and

the City of Vancouver By-Law no. 9344.

Environmental noise, unlike many other types of pollution, is dependent on local

physical conditions. Many factors including development planning, the structure of the built

environment, population density, as well as local habits and culture can determine how

dangerous environmental noise is to the local population (25). The World Health Organization

(WHO) has set maximum values for environmental noise with respect to human health but

these do not take into account local factors which can affect environmental noise.

3.1 Sampling Strategies

Ambient environmental noise studies assess noise within a defined spatial region, such

as a city, and describe the cumulative exposure to residents. Three general strategies for

investigating environmental noise in an urban environment include sampling by grid, road

classification, and studies that are restricted to either a source or receptor of environmental

noise (26). The grid sampling method is the system currently recommended by the

International Standards Organization (ISO) for ambient environmental noise studies (27). This

Page | 11

method uses an evenly distributed sampling grid, resulting in good spatial distribution data.

The classification sampling system uses road type or land use classes to break a city up into

study areas. Average noise values are assigned to all areas of that classification type.

Classification studies are the most economical in terms of cost and labour and present data in

an easy to understand format. Studies of these types use a few study points to characterize

each identified class, generating a set of average values which can be applied for all areas of

that type and are generally used for creating thematic noise maps. These advantages make grid

sampling and classification studies the ones most commonly used by cities and noise consulting

firms.

Source studies examine either a physical point source of environmental noise, like an

airport (28) or a vector source, like a road network (29). Source studies focus on the planning

and development of new infrastructure to limit the exposure of nearby residents to harmful

levels of sound, or on planning noise abatement procedures (23;29). Source studies are

conducted by examining the peak levels of sound produced by the target source and comparing

them to appropriate guidelines.

Receptor study designs examine the effect of environmental noise on a target receptor

by studying multiple cohorts exposed to different amounts of environmental noise (5). These

designs usually perform an assessment of the exposure levels of a target group. Health

researchers commonly use this type of study to examine the long-term impacts of noise on

human health (29).

Page | 12

3.2 Sampling Periods

The written literature has little agreement on the appropriate frequency and duration to

sample environmental noise. Sampling frequencies and approaches vary greatly among

studies, and include 15-minute measurements every two hours (22), daytime-only assessment

(25), day and night measurements(30), and continuous assessment (31). Sommerhoff et al. (24)

and Ng et al. (32) used a three-period assessment, dividing the 24-hour clock into three periods

(day, evening and night) but differed slightly in their period start times and sample lengths.

This study used a modified version of the three-period assessment method, and incorporated

refinements discussed by Ng et al. (32) for improving the statistical accuracy of the testing.

A three-period division of the day may not be sufficiently accurate for a full assessment

of the quantity of environmental noise produced in a day. Daytime periods should be

subdivided into two sections, morning and afternoon, resulting in a statistically valuable

improvement in assessment quality (32). Moreover, only the first or last two hours of the night

period should be sampled so as to include the times of highest environmental noise production.

Dividing the night periods into two sample periods has not been shown to have a statistically-

significant effect on the quality of the data, and may increase resident disturbance (32).

Page | 13

3.3 Health Related Values

Health standards for environmental noise are developed by various organizations

including WHO, the (European Union) EU and the US Environmental Protection Agency (EPA) as

well as by specific countries or cities. These values consist of sound levels that are acceptable

during parts of the day commonly divided by the dominate usage of the area. Values are based

primarily on sleep disturbance and acceptable annoyance levels (33;34).

The EPA uses two time periods, day (07:00-22:00) and night (22:00-07:00) to define

allowable environmental noise levels and classes of land use in an effort to provide guidelines

that reflect usage and the local population base. Urban residential areas have a daytime

recommended maximum exposure of 52 dB (A) and a night-time value of 45 dB (A) while a

predominantly industrial area has a limit of 70 dB (A) for both night and day (table 1) (33). Italy

is one of the few countries that have put noise exposure limits into legislation. They have used

the same general procedure as the EPA and defined day (06:00 – 22:00) and night (22:00-06:00)

periods which have permissible limits for the equivalent continuous sound level in dB (A).

Residential areas in the Italian legislation are limited to 55 dB (A) and 45 dB (A) at night, while

exclusively industrial areas are limited to 70 dB (A) for both the day and night (table 2) (22).

Page | 14

Land use type Maximum noise level guidelines dB (A)

Time period 07:00 – 22:00 22:00 – 07:00

Rural residential 47 40

Urban residential 52 45

Urban residential with light commerce or institutional

55 45

Urban residential with light manufacturing, public

entertainment or licensed premises

58 50

Commercial 65 60

Industrial 70 70

Table 1. EPA guidelines for environmental noise as defined by land use type and time of day.

Page | 15

Land use type Maximum noise level guidelines dB (A)

Time period 06:00 -22:00 22:00 – 06:00

Protected areas 50 40

Residential areas 55 45

Mixed areas 60 50

Areas of intense activity 65 55

Predominantly industrial

areas

70 60

Exclusively industrial areas 70 70

Table 2. Acceptable levels of environmental noise as defined by Italian Legislation for times of day and land use type.

Page | 16

3.4 Data Analysis

Data analysis is typically conducted using descriptive statistical methods. The simplest

of these is where the maximum (LAmax) and minimum (LAmin) noise levels are the highest and

lowest levels observed respectively. Other common statistical descriptors include LA1, LA5, and

LA90 and are used to describe the noise in a specified percentile of the sample period. LA1 and

LA5 measure the highest levels of sound and are commonly used to assess traffic noises while

LA90 is used to examine background environmental noise. As statistical descriptors can be

cumbersome, the equivalent sound level (Leq) is used to describe the average sound level during

a stated period of time and is used as a common substitute. Composite whole–day ratings are

also used to describe environmental noise and two ratings are commonly employed. Day-Night

Level (Ldn) uses two Leq values - one for day and one for night while Day-Evening-Night (LRden)

uses three Leq values and produces a more accurate reflection of ambient environmental noise.

All of these statistical descriptors are in common use and calculation methods are described in

ISO 1996-1 (25;32;35). Results are usually presented as thematic maps to display spatial

variation while line graphs are used for temporal variation (23). Tables, box-and-whisker

diagrams and histograms are frequently used to compare between different testing locations

(25).

Page | 17

4. METHODS

Two neighbourhoods on the Halifax peninsula area were identified as representative of

residential and mixed use environments in urban Halifax. These areas were selected from

regions delineated by the Federal Government as dissemination areas, part of the national

census program; dissemination areas are the smallest units used by the census program.

Several of these areas make up each census tract, resulting in tracts with similar population

sizes, regardless of the physical area defined by the boundaries.

The criteria used for selecting study areas were the structure of the local built

environment, the population density, the geographic region, the surrounding area, and the

local land-use using data from the 2001 Canadian census and DMTI Spatial(36). All geographic

information data (GIS) used, including census information, city road maps and land-use are

available from the Maps & Geospatial Information Collection (MAGIC) administered by the

Dalhousie GIS Centre.

4.1 Study areas

The chosen dissemination areas identified for the study are located in the south end of

Halifax (figure 1). Area one, the representative residential area, is bound by Robie Street in the

east, Bellevue and Waterloo Streets to the west, and South and Roxton Streets in the north and

south respectively, giving it a north - south oriented rectangular shape. This area is

predominantly composed of single-family dwellings, two to three stories in height, and has a

total population of 653 permanent residents. Buildings in this area are free standing and

Page | 18

constructed wood, stone and brick. The total area of the Area one is 0.16 Km2 and it has a

population density of 3967.2 per km2. There are a total of seven streets inside the boundaries

of the region and a length of 3506 m for all roads contained in the sample region.

Area two, the mixed use area, has a larger area of 0.30 km2, but a smaller population of

566, resulting in a population density of 1836.5 per Km2. This area is bound by Hollis and

Barrington Streets in the east, South Park, Brenton and Queen Streets in the west, Sackville

Street to the north, and Spring Garden, Clyde, Morris, and Bishop Streets in the south. Area

two does not have a regular shape but is generally oriented east-west. This area contains

commercial, institutional, and residential zones, with multi-story buildings primarily build out of

concrete. There are 16 internal streets with a total road length of 6271 m.

Page | 19

Figure 1. Map of study areas with Area 1 in red and Area 2 in blue, sample sites are also identified.

Page | 20

4.2 Instrumentation

Instrumentation used in this study includes a Centre 322 Logging Sound Level Meter

(SLM) and a Marantz PMD-660 Solid State Digital Recorder (figure 2). The Centre SLM is an

ANSI S1.4 Type 2 instrument with a 0.5” electrets condenser microphone, and with a frequency

range of 31.5 Hz to 8 KHz, and a measuring level range of 30 to 130 dB. This instrument can

weight frequencies to either the A or C scale, and adjust the time weighting for either fast (125

ms) or slow (1 sec ) response. Under reference conditions, the SLM has an accuracy of +/- 1.5

dB at 94 dB, 1 KHz and can be manually calibrated. Logging capabilities of the unit include an

onboard memory of 32 000 data points and it can be connected to a Windows-based computer

Figure 2. Center 322 SLM (white) and Marantz 660 recorder (black) with microphone, wrist watch for scale.

Page | 21

via a RS232 port for either direct logging or to download stored information. The

accompanying software allows the collected data to be displayed in either graphical or tabular

format(37).

The Marantz PMD-660 Solid State Digital Recorder has a sampling frequency of 44.1, 48

KHz with a frequency response of 16,000 Hz. The recorder is connected to an external

microphone and it can record 4 hours of data at the stated frequency, which can be

downloaded to a computer for analysis (38).

4.3 Sampling periods

The study was conducted using a 45-minute sampling period for all samples.

Each sample location was characterized by three hours of data distributed over each of the four

time divisions identified in table one. The hour in which the sample was conducted was

randomly assigned to different sample points but in a manner that insured the full six hour

period was sampled. Each sample was started at the top of the hour. Sample point locations

are based on the sample grid system identified in ISO standard 1996-2 as being the most

effective manner of assessing environmental noise with one randomly located sample point in

each grid square.

4.4 Sample Points

The location of the random sample points was determined by GIS manipulation of the

Halifax city roads map file in ArcGIS. A four meter buffer was created inside the curb, away

Page | 22

from the roads, inside the study area to create the sample area. To do this a new feature layer

is created in ArcCatalog, and the curbs of all streets inside the study areas are traced by using

the editor tool. A four meter wide buffer is then generated by ArcGIS on the internal side of the

curb of the study area using the buffer tool. The study areas were divided into six sample

squares defined by the size of the sample area, the roads buffer polygon is extracted from

inside each grid square. A random point is then generated by Hawth’s Analysis Tools for ArcGIS

inside the buffered area. This produces a grid sampling plan with a randomly identified point in

each grid square in the four meter buffered region from each road.

Measurements collected in this study were taken at a height of 1.5 m, at a distance of

0.5 m from the curb, with the SLM oriented perpendicular to the nearest road. The SLM and

sound recorder were mounted on a camera tripod and a microphone stand which were locked

in place. The SLM averages one second measurements while the sound recorder produced a

continuous record. Measurements were not taken on days with rain, snow or high winds,

because these elements can both damage equipment and decrease the accuracy of the

measurements.

4.5 Analysis

The SLM data provides the minimum and maximum sound pressure level (SPL)

averaged over one second, resulting in 2700 data points for each sample, and 10800 data

points for each grid sample area in a 24-hour period. The sound recorder provided continuous

recording which was used to identify peak noise events. Basic noise descriptors, including

Page | 23

maximum and minimum, mean, and various percentile sound levels which are used to describe

the level of sound equalled or exceeded during the identified percentage of a time period.

Also, equivalent continuous sound pressure levels (Leq) and day – evening – night composite

whole-day rating level (Lrden) descriptors were then produced for the sample periods, grid

sample areas and study areas to identify variations in environmental noise over both space and

time.

The two study areas were compared both spatially and temporally using statistical

analysis of the environmental noise. Each study area was examined individually to determine

spatial variation of environmental noise during each six hour sample period, as well as during

the full 24-hour period. Comparative analysis between the individual sample sites was

conducted using Kruskal – Wallis tests while analysis between the two study areas was

conducted using the Mann-Whitney two sample rank test and Leven’s test.

The Kruskal–Wallis test is used when there is one nominal variable and one

measurement variable which does not meet the normality assumption of ANOVA. A one-way

ANOVA may yield inaccurate estimates of the P-value when the data are very far from normally

distributed. The Kruskal–Wallis test does not make assumptions about normality; it is

performed on ranked data, so the measurement observations are converted to their ranks in

the overall data set. The Mann-Whitney U test is a non-parametric test for assessing whether

two samples of observations come from the same distribution. The null hypothesis is that the

two samples are drawn from a single population, and therefore that their probability

distributions are equal (39). The Levene's test tests the null hypothesis that the population

http://udel.edu/~mcdonald/statvartypes.html#nominal

http://udel.edu/~mcdonald/statvartypes.html#measurement

http://udel.edu/~mcdonald/statnormal.html

http://udel.edu/~mcdonald/statvartypes.html

http://en.wikipedia.org/wiki/Non-parametric_statistics

http://en.wikipedia.org/wiki/Sampling_%28statistics%29

http://en.wikipedia.org/wiki/Null_hypothesis

http://en.wikipedia.org/wiki/Probability_distribution



Page | 24

variances are equal. If the resulting value is significant at the identified level the variance

between the two populations is unlikely to have occurred based on random sampling (40).

4.6 Health Comparisons

By using the available acceptable levels of environmental noise from the EPA and

Italian legislation for the appropriate land use types, a base was established for comparison

purposes. The data was then recalculated to conform to the standardized time periods. The

two sample areas were compared to the established guidelines using both un-weighted and

weighted LAeq values, to include disturbance impact of time of day, to examine the acceptability

levels of environmental noise and to assess the health risks to the local population.

Page | 25

5. DATA PRESENTATION AND ANALYSIS

Data recorded during this study includes 36 hours of both sound recording and SLM data

points. Each of the areas are characterized by 18 hours of data with each site having three

hours, with four samples of 45 minutes providing 2700 data points for each 45 minute sample.

For each sample a LAeq has been calculated using formula (1) producing the A scale weighted

equivalent continuous sound pressure level as described in the ISO 1996-1 standards (35).

dtp

tPT

Lo

a

Aeq 2

21log10

(1)

Where:

PA2 is the A-weighted instantaneous sound pressure at the running time t;

po is the standard reference sound 20 µPa;

The LAeq values were then adjusted according to Annex A of the ISO 1996-1

standards to reflect the time period in which the sample was taken. Annex A allows an

adjustment of +5 dB for the evening hours and +10 for the night hours to reflect the

disturbance of the sounds using formula (2) to produce and adjusted LAeq.

jTnAeqjTnqj KLL ,,Re (2)

Where:

Kj Is the adjustment for the specified sample and time period;

LAeqj,Tn Is the actual LAeq at the specified time period;

Using the adjusted LAeq values the composite whole-day rating level was then

calculated to develop a day-evening-night rating level using equation (3).

Page | 26

dbx

edx

ex

dL

LRnLLRD

Rden

1010

Re

10 1024

2410

2410

24log10 (3)

Where:

d is the number of daytime hours;

n is the number of night-time hours;

e is the number of evening hours;

LRd is the rating level for daytime hours including adjustments;

LRe is the rating level for evening hours including adjustments;

LRn is the rating level for night-time hours including adjustments;

5.1 Area one

Area one sample points show variation between both individual sites and sample

periods. Maximum values for the individual sites ranged from 63.7 dB (A) to 93.3 dB (A) with

most site values in the 70’s. LA90 values for the sample sites range from a low of 38.2 dB (A) to a

high of 50 dB (A) and represent the background noise levels of the area. Non-adjusted LAeq

values range from 42.7 dB (A) to 69.1 dB (A) with the difference between the maximum and

minimum LAeq values for all sites between 9.6 and 13.4 (table 3). When the data points for area

one are grouped into the four sample time periods and averaged the values for the maximum

SPL levels are between 71.3 dB (A) and 77.4 dB (A). LA90 values for the four time periods range

Area 1 Percentiles

Sample Site Period Start Time Maximum Min Mean LA1 LA5 LA10 LA90 LA95 LA99 LAeq Adjusted LAeq

1 1 7:00 73 40.1 44.2 60 51.1 46.9 41.8 41.4 40.6 48.5 48.5

2 16:00 73.3 41.4 47.7 65.1 59.1 56.3 42.8 42.4 41.8 53.2 53.2

3 18:00 66.6 25.8 43.9 61.8 55 51.5 39.5 38.9 38.4 49.2 54.2

4 3:00 66.3 41.7 43.9 51.6 45.5 44.9 42.8 42.6 42 45 55

Maximum LAeq 53.2 55

Minimum LAeq 45 48.5

Difference 8.2 6.5

2 1 8:00 72.9 43.7 51.3 67.4 60.6 57 46.3 45.6 44 55.4 55.4

2 12:00 75.4 40.9 48 63.7 56.2 53.3 43.3 42.9 43.3 53 53

3 22:00 65.2 21 44.2 55.9 51.5 49.1 41.5 41.2 40.8 46.6 51.6

4 1:00 66.3 38.8 40.3 49.2 42.2 40.9 39.4 39.3 39.2 42 52

Maximum LAeq 55.4 55.4

Minimum LAeq 42 51.6

Difference 13.4 3.8

3 1 9:00 90 42.3 61.4 80.3 73.4 71.8 48 46.1 43.9 69.1 69.1

2 14:00 86.6 40 57.7 76.3 72.5 70.3 45.7 43.2 41.1 66.3 66.3

3 23:00 81.4 37 43.1 72.1 62.3 54.3 38.2 37.9 37.5 58.6 63.6

4 5:00 93.3 20 48 77.6 67.7 59.1 43.3 43.1 42.8 66.8 76.8


Minimum LAeq 58.6 63.6

Difference 10.5 13.2

Table 3. Summary table of various statistical values for Area 1.

Table 3. Summary table of various statistical values for Area 1.

Page | 2

7

Area 1 Percentiles


4 1 10:00 79.8 47 58.1 73.8 69.5 67.1 50.3 49.4 47.7 63.1 63.1

2 15:00 77.5 23 49 66.7 59 56.7 43 42.3 41.5 54.8 54.8

3 21:00 78.9 43.9 53.8 71.3 66.6 63.7 46.8 46.1 44.8 60 65

4 24:00 77.4 39.9 45.8 63.5 58.1 55 41.3 41 40.6 52.9 62.9

Maximum LAeq 63.1 65


Difference 10.2 8.3

5 1 11:00 72.7 41.6 50.8 66.6 60.5 58.8 43.9 43.2 42.3 55.4 55.4

2 13:00 77.5 23 49 66.7 59 56.7 43 42.3 41.5 54.8 54.8

3 19:00 73.9 37.8 48.4 65 59.8 57.1 40.5 39.8 38.8 54.2 59.2

4 4:00 63.7 42.2 44.4 53.4 46.9 45.7 43 42.8 42.4 45.2 55.2



Difference 9.6 4.4

6 1 6:00 67.9 40.5 43.1 50 45.6 44.3 41.9 41.7 41.3 44.7 44.7

2 17:00 73.8 42.6 49.8 66.5 59.6 56.4 45.6 45 43.9 54.6 54.6

3 20:00 73.4 41.2 45.5 61.7 51.6 48.9 43 42.6 42 50.3 55.3

4 2:00 60.6 38.3 41.7 51.5 44.6 42.8 40.2 39.7 38.8 42.7 52.7




Table 3 (Continued). Continuation of table three showing statistical values for Area 1.

Page | 2

8

Page | 29

from a low of 41.6 dB (A) to a high of 45.4 dB (A) while non-adjusted LAeq values range from a

low of 49.1 dB (A) to a high of 56 dB (A) (table 4).

Area one shows a highly variable amount of sound both spatially and temporally. A few

sample sites have statistically greater levels of environmental noise as well as significant peaks

in the maximum SPL level. Sites three and four show higher than average LA1 (figure 3) and LA5

SPL levels, indicating high levels of road traffic near the sample points. The LA90 levels

representing the background levels of environmental noise are more consistent with other

sample points in Area one (figure 4).

Figure 3. Area 1 LA1 values for the four sample time periods, LA1 is a measure of environmental noise from traffic

Figure 4. Area 1 LA90 values which represent the background environmental noise levels.

Page | 30

Area 1 Divided by Period

Morning Percentiles

Site Time Maximum Mean LA1 LA10 LA90 LAeq Adjusted LAeq

6 6:00 67.3 43.1 50 44.3 41.9 44.7 44.7

1 7:00 73 44.2 60 46.9 41.8 48.5 48.5

2 8:00 72.9 51.3 67.4 57 46.3 55.4 55.4

3 9:00 90 61.4 80.3 71.8 48 69.1 69.1

4 10:00 79.8 58.1 73.8 67.1 50.3 63.1 63.1

5 11:00 72.7 50.8 66.6 58.8 43.9 55.4 55.4

Average 76.0 51.5 66.4 57.7 45.4 56.0 56.0

Afternoon

2 12:00 75.4 48 63.7 53.3 43.3 53 53

5 13:00 77.5 49 66.7 56.7 43 54.8 54.8

3 14:00 86.6 57.7 76.3 70.3 45.7 66.3 66.3

4 15:00 77.5 49 66.7 56.7 43 54.8 54.8

1 16:00 73.3 47.7 65.1 56.3 42.8 53.2 53.2

6 17:00 73.8 49.8 66.5 56.4 45.6 54.6 54.6

Average 77.4 50.2 67.5 58.3 43.9 56.1 56.1

Evening

1 18:00 66.6 43.9 61.8 51.5 39.5 49.2 54.2

5 19:00 73.9 48.4 65 57.1 40.5 54.2 59.2

6 20:00 73.4 45.5 61.7 48.9 43 50.3 55.3

4 21:00 78.9 53.8 71.3 63.7 46.8 60 65

2 22:00 65.2 44.2 55.9 49.1 41.5 46.6 51.6

3 23:00 81.4 43.1 72.1 54.3 38.2 58.6 63.6

Average 73.2 46.5 64.6 54.1 41.6 53.2 58.2

Night

4 24:00 77.4 45.8 63.5 55 41.3 52.9 62.9

2 1:00 66.3 40.3 49.2 40.9 39.4 42 52

6 2:00 60.6 41.7 51.5 42.8 40.2 42.7 52.7

1 3:00 66.3 43.9 51.6 44.9 42.8 45 55

5 4:00 63.7 44.4 53.4 45.7 43 45.2 55.2

3 5:00 93.3 48 77.6 59.1 43.3 66.8 76.8

Average 71.3 44.0 57.8 48.1 41.7 49.1 59.1

Maximum mean 1% 90% Laeq Adjusted LAeq

morning 76 51.5 66.4 45.4 56 56.0

afternoon 77.4 50.2 67.5 43.9 56.1 56.1

evening 73.2 46.5 64.6 41.6 53.2 58.2

night 71.3 44 57.8 41.7 49.1 59.1

Table 4. Statistical values for Area 1 distributed by sample time period.

Page | 31

When LAeq values for the full day are plotted, there are anomalous LAeq peaks located at

05:00, 09:00, 14:00, 21:00, 23:00 and 00:00 with a smaller peak at 19:00 (figure 5). Data for

05:00, 09:00, 14:00, and 23:00 is from site three while 21:00 and 24:00 is from site four and the

smaller peak at 19:00 is data from site five demonstrating the impact of higher volumes of

traffic noise on environmental noise levels. Non-adjusted LAeq values reflect the same pattern as

the maximum and road traffic values with sites three and four pulling the study area average

upwards. Using equation (3) the composite whole day rating for Area one is 63.8 dB (A)

Figure 5. Area 1 full 24-hour LAeq values with adjustment.

Page | 32

A histogram plot of all data points recorded in Area one demonstrates a skewed right

normal distribution with a mean dB (A) of 48.05 and a standard deviation of 7.06 demonstrating

a minimum level of environmental noise and the significance of traffic noise on the data point

distribution (Figure 6).

Individual sample sites vary greatly in data point distribution both between sites and

between sample periods as the same site. Boxplots and interval plots (CI, 95%) reveal

differences in the structure of specific site data between sites with more or less environmental

noise. Site 6, exhibits the lowest level of environmental noise in the area. This site also

presents a low IRQ box with significant numbers of high outlier values throughout the day

which correspond to elevated noise events that could cause disturbance. In night and early

morning samples low outliers also occur which are moments of very low amounts of

environmental noise. Standard deviation from the mean averaged for all sample periods in site

85.576.066.557.047.538.028.519.0

4000

3000

2000

1000

0

Area 1

Freq

uenc

y

Mean 48.05

StDev 7.605

N 64824

Area 1 data point distribution

Figure 6. Histogram plot of all Area one data points illustrating a skewed right normal data distribution and offset fits line.

Page | 33

6 is 3.015 (figure 8).

Sample site 3 which had the highest levels of environmental noise in Area 1

demonstrates a different data distribution pattern. Site 3 exhibits higher IRQ boxes and fewer

outlier points, all of which are in the evening and night sample time periods. During the day

there is a steady background level of traffic absent from sites more distant from high traffic

roads and the traffic events blend into the background noise (figure 9). Average standard

deviation from the mean for site 3 is 8.546.

To examine if the individual sample sites in Area one have statistically different levels of

environmental noise a Kruskal - Wallis Test was conducted (figure 7). This test presents

sufficient evidence to reject the null hypothesis that all median values are equal in favour of the

alternative hypothesis which states at least one is not equal in terms of environmental noise

level at a significance level of 95%.

Kruskal-Wallis Test – area 1

Test Statisticsa,b

16.232

5

.006

Chi-Square

df

Asy mp. Sig.

adjlaeq

Kruskal Wallis Testa.

Grouping Variable: Sample Siteb.

Figure 7. Kruskal-Wallis test for area 1 sample sites exhibiting sufficient evidence to reject the null hypotheses.

Page | 34

Figure 8. Boxplot and 95% mean confidence interval for site 6, Area 1. Sample start time on the horizontal axis with SPL in dB (A) on the vertical axis.

Page | 35

Figure 9. Boxplot and 95% mean confidence interval for site 6, Area 1. Sample start time on the horizontal axis with SPL in dB (A) on the vertical axis.

Page | 36

5.2 Area two

Sample points in Area two exhibit less variation between recorded values for individual

sites and the sample periods than Area one. Peak SPL levels range from 90.3 dB (A) to 69.7 dB

(A) while LA90 values range from a high of 59.3 dB (A) to a low of 44 dB (A). Area non adjusted

LAeq values range from 71.1 dB (A) to 49.6 db (A) while the difference between the maximum

and minimum LAeq levels range from 15.4 dB (A) and 5.1 dB (A) (table 5). Time period data for

Area two presents average maximum SPL levels between 84.9 dB (A) and 77.2 dB (A) while LA90

averaged values range from 54.6 dB (A) and 47.1 dB (A). Non-adjusted LAeq for Area two time

periods range from 64.1 dB (A) to 56.1 dB (A) (table 6).

Area two, the mixed use area, demonstrates a more consistent level of environmental

noise across all the sample sites. Over the four time periods maximum SPL as well as LA1 (figure

10) and LA5 are similar for all sites demonstrating a consistent level of vehicular noise

throughout the area over the full day. LA90 is highest in the afternoon at 54.6 dB (A), similar to

the morning reading of 53.1 dB (A) and decreases through the evening to 47.1 dB (A) at night

(figure 11).

Page | 37

Area 2 Percentiles


1 1 9:00 87 52.3 63.1 79.1 74.1 70.5 56.4 55.2 53.4 68.2 68.2

2 12:00 88.3 55.4 65.1 75.9 71.6 70.1 59.3 58.4 57.3 58.1 58.1

3 20:00 77.3 49.1 56 69.1 64.5 61.6 51.4 50.7 49.8 59 64

4 2:00 79.4 42.3 50 65.3 59.7 56.7 45.9 45.1 44.1 55.8 65.8




2 1 8:00 89 46.7 58.3 75.1 71.8 67.4 52.2 51.1 49.2 65 65

2 14:00 85.9 46.7 56 69.3 63.9 61.1 51.9 51.2 50.2 60.8 60.8

3 23:00 77.8 48.9 53.4 67.3 60.5 58.3 50.2 49.8 49.4 56.7 61.7

4 4:00 69.7 42.5 47.3 59.9 53.3 50.7 44.9 44.4 43.5 49.6 59.6

Maximum LAeq 65 65


Difference 15.4 5.4

3 1 10:00 86.8 54.5 60.8 77 73.4 67.8 56.2 55.7 55 66 66

2 15:00 85.2 54.3 60.3 71.6 66.9 65.1 56.6 56.2 55.6 62.7 62.7

3 18:00 83.3 54.1 60.4 72.5 68.4 66.6 55.9 55.6 54.9 63.5 68.3

4 5:00 75.1 49.7 54 67.5 60.2 58 51.5 51 50.4 56.4 66.4

Maximum LAeq 66 68.3


Difference 9.6 5.6

Table 5. Summary table of various statistical values for Area 2

Page | 3

7

Page | 38

Area 2 Percentiles


4 1 11:00 72.7 45.4 52.6 65.9 60.8 58.1 49 48.3 46.9 55.4 55.4

2 13:00 83.4 47.3 53.7 67.7 62 59.4 50 49.4 48.4 58.5 58.5

3 22:00 75.1 28.6 50.4 66 58.6 55.3 47.2 46.8 46.2 54.1 59.1

4 24:00 71.9 45.7 49.7 62.9 56.2 53.4 47.3 47 46.4 52.4 62.4



Difference 6.1 7

5 1 6:00 77.3 47.4 54 70.1 65.4 62.2 49 48.6 48.1 58.9 58.9

2 16:00 86 23.7 60.9 72.1 69.1 67.1 55 53.7 51.8 64 64

3 19:00 77.6 48.5 57.7 72.3 68.4 66.4 51.7 50.9 49.9 62.1 67.1

4 1:00 85.7 46.2 53.8 73.5 65.9 61.4 49.1 48.7 48.1 61.3 71.3

Maximum LAeq 64 71.3


Difference 5.1 12.4

6 1 7:00 90.3 49.8 65.6 81.3 76.5 74.3 56 54.2 52 71.1 71.1

2 17:00 80.4 49.6 63.1 75.7 71.9 70.2 54.9 53.6 51 66.7 66.7

3 21:00 83.7 46.7 60.1 74.3 70.6 68.6 51.8 50.5 48.5 64.8 69.4

4 3:00 81.4 23.6 51.7 75.6 68.6 63.6 44 43.6 43 62.2 72.2



Difference 8.9 5.5

Table 5 (continuation). Continuation of summary data table of various statistical values for Area 2

Page | 3

8

Page | 39

Area 2 Divided by Period

Morning

Site Time Maximum Mean LA1 LA10 LA90 LAeq Adjusted LAeq

5 6:00 77.3 54 70.1 62.2 49 58.9 58.9

6 7:00 90.3 65.6 81.3 74.3 56 71.1 71.1

2 8:00 89 58.3 75.1 67.4 52.2 65 65

1 9:00 87 63.1 79.1 70.5 56.4 68.2 68.2

3 10:00 86.8 60.8 77 67.8 56.2 66 66

4 11:00 72.7 52.6 65.9 58.1 49 55.39 55.4

Average 83.9 59.1 74.8 66.7 53.1 64.1 64.1

Afternoon

1 12:00 88.3 65.1 75.9 70.1 59.3 58.1 58.1

4 13:00 83.4 53.7 67.7 59.4 50 58.5 58.5

2 14:00 85.9 56 69.3 61.1 51.9 60.8 60.8

3 15:00 85.2 60.3 71.6 65.1 56.6 62.7 62.7

5 16:00 86 60.9 72.1 67.1 55 64 64

6 17:00 80.4 63.1 75.7 70.2 54.9 66.7 66.7

Average 84.9 59.9 72.1 65.5 54.6 61.8 61.8

Evening

3 18:00 83.3 60.4 72.5 66.6 55.9 63.5 68.5

5 19:00 77.6 57.7 72.3 66.4 51.7 62.1 67.1

1 20:00 77.3 56 69.1 61.6 51.4 59 64

6 21:00 83.7 60.1 74.3 68.6 51.8 64.8 69.4

4 22:00 75.1 50.4 66 55.3 47.2 54.1 59.1

2 23:00 77.8 53.4 67.3 58.3 50.2 56.7 61.7

Average 79.1 56.3 70.3 62.8 51.4 60.0 65.0

Night

4 24:00 71.9 49.7 62.9 53.4 47.3 52.4 62.4

5 1:00 85.7 53.8 73.5 61.4 49.1 61.3 71.3

1 2:00 79.4 50 65.3 56.7 45.9 55.8 65.8

6 3:00 81.4 51.7 75.6 63.6 44 62.2 72.2

2 4:00 69.7 47.3 59.9 50.7 44.9 49.6 59.6

3 5:00 75.1 54 67.5 58 51.5 56.4 66.4

Average 77.2 51.1 67.5 57.3 47.1 56.3 66.3

Maximum mean LA1 LAeq LA90 Adjusted LAeq

morning 83.9 59.1 74.8 64.1 53.1 64.1

afternoon 84.9 59.9 72.1 61.8 54.6 61.8

evening 79.1 56.3 70.3 60.0 51.4 65.0

night 77.2 51.1 67.5 56.3 47.1 66.3

Table 6. Area 2 selected statistical values divided by sample period.

Page | 40

Figure 10. LA1 values from Area 2, representing traffic induced environmental noise.

Figure 11. LA90 values for Area 2 by site measuring background environmental noise.

Page | 41

The twenty-four hour plot of LAeq for Area two demonstrates anomalous peaks at 01:00, 03:00,

07:00 and 21:00. The last three LAeq values are data from site six while the 01:00 peak is data

from site 5 (figure 12). Site six is located on Barrington Street which is a high traffic street with

significant bus traffic resulting in increased environmental noise.

Area two demonstrates a similar skewed right data distribution plot as Area one with a

mean of 56.59 dB (A) and a standard deviation of 7.088. The data is more evenly distributed

through the curve and the distribution is more symmetric (figure 13). The composite whole day

rating as calculated by equation (3) gives a result of 65 dB (A).

Figure 12. Area 2 LAeq and adjusted LAeq values for the full day.

Page | 42

Area two sample sites exhibited similar patterns of variation between sample sites and

times as Area one. Sample sites near high traffic roads exhibited higher IRQ boxes and fewer

outlier points (figure 14). Area two and fewer outlier points than Area one, resulting from

overall higher levels of environmental noise, traffic events tend to blend into the background

noise and do not cause significant increases in sound. Sample sites that have less road traffic

exhibited lower IRQ boxes and more outlier points, resulting from lower ambient levels of noise

(figure 15). For Area two sample sites a Kruskal – Wallis test produces a similar result as that

for Area one. There is sufficient evidence to reject the null hypothesis that all sample sites have

an equal median amount of environmental noise in favour of the alterative that they do not.

9081726354453627

2000

1500

1000

500

0

Area 2

Freq

uenc

y

Mean 56.59

StDev 7.088

N 64824

Area two data distribution

Figure 13. Distribution histogram of all data points in Area 2 and fits line.

Page | 43

Figure 14. Boxplot and 95% confidence interval for sample site 6 Area 2. Sample start time on the horizontal axis and SPL level in dB (A) on the vertical axis.

Page | 44

Figure 15. Boxplot and 95% confidence interval for sample site 3 Area 2. Sample start time on the horizontal axis and SPL level in dB (A) on the vertical axis.

Page | 45

5.3 Analysis

There are differences between the two sample areas, both in terms of distribution of

sound and overall levels of environmental noise. Adjusted LAeq values are more variable

between sample sites in Area one than sample sites in Area two (figure 16). The difference

between the sites is an artefact of the variability of traffic volumes related to land use,

background institutional noise and pedestrian activity. The noisier sites in Area one are those

near major roads while the quieter sites are the most distant from the same roads. Area two

has higher levels of environmental noise with more consistency between the sample sites.

There is some variability between the sites but due to the overall greater level of vehicular

traffic in the overall area the inconsistency is lower. There is significantly more institutional and

Figure 16. Area 1 & 2 adjusted LAeq values for all sites. Area 1 demonstrates much more variability between sample sites.

Page | 46

industrial noise in Area two as a result of the land use. This includes ventilation systems,

delivery vehicles and increased levels of pedestrian activity.

The LA90 level of environmental noise is lower in Area one than Area two. Area two has

more background noise throughout the day resulting from vehicle traffic in the area, industrial

noises like ventilation fans, delivery trucks and high pedestrian traffic. Similarly the adjusted

LAeq values for Area one are lower than those in Area two as a result of the land use (figure 17).

Area one is more vulnerable to the disturbance effects of noise events. A vehicle

passing through Area one could cause a increase of sound of 10 to 30 dB (A) causing residence

disturbance while in Area two the same vehicle may either be lost in the higher background

level of sound or increase the levels of sound by a smaller amount. The composite full day

rating (LRden) values for the two areas show very little difference in daily sound exposure. Area

one is 63.8 dB (A) and Area two is 65 db (A).

Figure 17. Areas 1 & 2 LA90 and Adjusted LAeq values area 2 values are approximately 10 dB (A) higher.

Page | 47

Using the EPA and Italian environmental noise guidelines for human health, Area one

adjusted and non-adjusted LAeq values exceed both sets of guideline values by several decibels.

Area two unadjusted LAeq values are considered acceptable under EPA guidelines but are

unacceptable during the night for the Italian guidelines. Adjusted LAeq values are unacceptable

under both sets of guidelines for both periods of the day (table 7).

Statistical analysis of the individual sample sites using Kruskal – Wallis tests provides

evidence that in both sample areas there is a statistically different level of environmental noise

either between the sample sites, or the sample time periods in either area. By conducting a

Mann-Whitney analysis on the difference between the median values for the two sample areas,

the test is significant at a value of 0.0002. This provides enough evidence to reject the null

hypotheses that the median values of the sample areas are equal and identifies a statistically

significant difference in the amount of environmental noise between the two study areas. This

statistical evidence supports the hypothesis that the built environment affects the level of

environmental noise to which residents are exposed. A Levene’s test of equality of error

variances also supports this hypothesis, by showing that both sample areas and sample sites

are significant factors in the variance with an R squared value of 0.477 (figure 18).

Page | 48

Health values

Area 1 Area 1 Health values

Area 2 Area 2

unadjusted LAeq

adjusted LAeq

unadjusted LAeq

adjusted LAeq

Urban Residential Predominantly commercial

EPA Values

Day 07:00-22:00

52 60.8 61.5 65 64.9 66.2

Night 22:00-07:00

45 58.2 66.7 60 57.9 66.9

Residential Areas Areas of intense activity

Italian Values

Day 06:00-22:00

55 60.5 61.2 65 64.5 65.9

Night 22:00-06:00

45 58.6 68.2 55 57.8 67.3

Table 7. LAeq and Adjusted LAeq values compared to EPA and Italian SPL levels permitted for health safety

Figure. 18 Levene’s test of equality of error variances demonstrating that both the study area and sample site are significant factors in the data point variance.

Tests of Between-Subjects Effects

Dependent Variable: adjlaeq

1068.259a 6 178.043 6.224 .000

177523.850 1 177523.850 6205.429 .000

576.160 1 576.160 20.140 .000

492.099 5 98.420 3.440 .011

1172.921 41 28.608

179765.030 48

2241.180 47

Source

Corrected Model

Intercept

area

SampleSite

Error

Total

Corrected Total

Type I II Sum

of Squares df Mean Square F Sig.

R Squared = .477 (Adjusted R Squared = .400)a.

Page | 49

6. LIMITATIONS, RECOMMENDATIONS AND CONCLUSIONS

The objective of this research is to observe and report on variations in environmental

noise with respect to built environments and temporal change. Two areas representative of

different types of land use were compared, looking at spatial variability of environmental noise

within and between sample areas. This study provides information regarding land use planning

decisions, the resulting pollution and the possible health implications for residents.

Study results indicate higher levels of environmental noise occur in mixed-use

neighbourhoods when compared to predominantly residential neighbourhoods. The variability

in environmental noise between the two study areas is a result of increased vehicular and

pedestrian traffic as well as background noise generated by institutional and industrial noise

like delivery trucks, and ventilation systems.

The variability of environmental noise is statistically significant within each sample area

between sample points, between the two study areas and temporally. Kruskal-Wallis tests

were used to examine differences in levels of environmental noise within the sample area and

presented enough evidence to reject the null hypotheses that all sites had equal amounts of

sound at a confidence level of 95%. As well, using a 2-sample Mann-Whittney test to examine

the variation in sound between the two sample areas also presented enough evidence to reject

the null hypotheses that the median values of the different areas at a confidence level of 95%.

Inside study Area one, environmental noise varied between sites as a result of traffic

patterns. Sites that were near high traffic roads with heavy truck or bus traffic exhibited higher

levels of environmental noise than more distant sites. Area one was also vulnerable to

Page | 50

disturbance as a result of traffic because of the lower levels of background noise allowing the

traffic to be more disruptive. Area two demonstrated less sample site variation in

environmental sound with sites near high traffic roads and bus routes recording higher levels of

sound. The variation was less than within Area one because there were higher overall levels of

traffic and greater background noise.

In Area one the absolute environmental noise levels were in the range of an office

environment or a normal conversation which is considered to be comfortable for humans. Area

two the absolute environmental noise levels were higher and can be considered intrusive for

normal conversation and slightly annoying. For both study areas the peak noise events ranged

from annoying to very annoying and obscured conversation.

Un-adjusted LAeq levels of environmental noise exceeded EPA and Italian guideline

values for human health impacts in Area one during the daytime. Area one exceeded the

Italian guideline values by 5.5 dB (A) during the day and 13.6 dB (A) and the EPA guidelines by

8.8 dB (A) during the day and 13.2 dB (A) at night. Area two Un-adjusted LAeq were acceptable

at all times except at night under the Italian guidelines where it was 2.8 dB (A) over guidelines.

Adjusted LAeq values were unacceptable by both sets of guidelines at all times.

Results of this study are not necessarily representative for all areas of the city as a result

of limitations placed on the study. Limitations include both financial and time constraints,

equipment used for this study was lower quality than international standards due to financial

constraints, and sampling areas and times were also limited to two areas and for a short

temporal frame as a result of man-power limitations. An increase in the number of study areas

Page | 51

across additional land-use types in Halifax would provide a more complete understanding of

the relationship between environmental noise, built environment and human health risks.

Sampling throughout the year would also reduce the influence of seasonal variation of

environmental noise levels as well full 24 hour samples would also remove measurement error

implicit in LAeq calculation.

Further study of environmental noise should be carried out in order to fully understand

the health risks to which residents of Halifax are exposed Future research should incorporate a

more comprehensive sampling strategy looking at different areas of Halifax and other land use

types and should be complimented by solicitation of perceived noise health impacts among

neighbourhood residents. Seasonal variation should also be examined as the source and

character of environmental noise may change due to weather and road condition changes.

Planning changes should also be undertaken to route heavy traffic to avoid residential

areas or sensitive areas like schools. City buses, a powerful noise source, should be required to

improve on muffler quality and Halifax should look to adopt environmental noise standards in

order to protect the health of both residents and the quality of the urban environment.

Page | 52

7. REFERENCE LIST

(1) Northridge ME, Sclar ED, Biswas P. Sorting out the connections between the built environment and health: a conceptual framework for navigating pathways and planning healthy cities. Journal of urban health 2003; 80(4):556.

(2) Jackson RJ. The impact of the built environment on health: an emerging field. Am J Public Health 2003; 93(9):1382-1384.

(3) Coensel BD, Muer TD, Yperman I, Botteldooren D. The influence of traffic flow dynamics on urban soundscapes. Applied Acoustics 2004; 66:175-194.

(4) Passchier-Vermeer W, Passchier WF. Noise Exposure and public health. Environmental Health Perspectives 2000; 108(1):123-131.

(5) Regecová V, Kellerová E. Effects of urban noise pollution on blood pressure and heart rate in preschool children. Journal of Hypertension 1995; 13:405-412.

(6) Stanfeld SA, Berglund B, Clark C, Lopez-Barrio I, Fischer P, Öhrström E et al. Aircraft and road traffic and children's cognition and health: a cross-national study. Lancet 2005; 365:1942-1949.

(7) Babisch W, Beule B, Schust M, Kersten N, Ising H. Traffic noise and risk of myocardial infarction. Epidemiology 2005; 16(1):33-40.

(8) Seto ETW, Holt A, Rivard T, Bhatia R. Spatial distribution of traffic induced noise exposures in a US city: an analytic tool for assessing the health impacts of urban planning decisions. International Journal of Health Geographics 2007.

(9) World Health Organization (WHO). Occupational and community noise. Fact scheet N.258 . 2001.

http://www.who.int/mediacentre/factsheets/fs258/en/print.html (10) Staff. Regional Municipal Planning Strategy. 2007. http://www.halifax.ca/regionalplanning/documents/Regional_MPS.pdf

(11) Zhang X, Chen J, Tan M, Sun Y. Assessing the impact of urban sprawl on soil resources of Nanjing city using satellite images and digital soil databases. CATENA 2007; 69(1):16-30.

Page | 53

(12) Bengston DN, Potts RS, Fan DP, Goetz EG. An analysis of the public discourse about urban sprawl in the United States: Monitoring concern about a major threat to forests. Forest Policy and Economics 2005; 7(5):745-756.

(13) Frenkel A. The potential effect of national growth-management policy on urban sprawl and the depletion of open spaces and farmland. Land Use Policy 2004; 21(4):357-369.

(14) Stone J. Urban sprawl and air quality in large US cities. Journal of Environmental Management 2008; 86(4):688-698.

(15) Wu J. Environmental amenities, urban sprawl, and community characteristics. Journal of Environmental Economics and Management 2006; 52(2):527-547.

(16) Couch C, Karecha J. Controlling urban sprawl: Some experiences from Liverpool. Cities 2006; 23(5):353-363.

(17) Eid J, Overman HG, Puga D, Turner MA. Fat city: Questioning the relationship between urban sprawl and obesity. Journal of Urban Economics 2008; 63(2):385-404.

(18) Hancock T. Health and sustainability in the urban environment. Environmental Impact Assessment Review 1996; 16(4-6):259-277.

(19) Mielke HW, Gonzales CR, Smith MK, Mielke PW. The Urban Environment and Children's Health: Soils as an Integrator of Lead, Zinc, and Cadmium in New Orleans, Louisiana, U.S.A. Environmental Research 1999; 81(2):117-129.

(20) de Hollander AEM, Staatsen BAM. Health, environment and quality of life: an epidemiological perspective on urban development. Landscape and Urban Planning 2003; 65(1-2):53-62.

(21) Mosskov JI, Ettema JH. Extra-auditory effects in long-term exposure to aircraft and traffic noise. Occupational and Environmental Health 1977; 40:177-184.

(22) Piccolo A, Plutino D, Cannistraro G. Evaluation and analysis of the environmental noise of Messina, Italy. Applied Acoustics 2005; 66:447-465.

(23) Arafa MH, Osmon TA, Abdel-Latif IA. Noise assessment and mitigation schemes for Hurghada airport. Applied Acoustics 2007; 68:1373-1385.

(24) Sommerhoff J, Recuero M, Suárez E. Community noise survey of the city of Valdivia, Chile. Applied Acoustics 2004; 65:643.

(25) Morillas JMB, Gómez Escobar V, Mѐndez Sierra JA, Vilchez Gómez R, Trujillo Carmona J. An environmental noise study in the city of Cáceres, Spain. Applied Acoustics 2002; 63:1061-1070.

Page | 54

(26) Brown AL, Lam KC. Urban noise surveys. Applied Acoustics 1987; 20:23-39.

(27) International Standards Organization. ISO 1996 - Acoustics - description, measurement and assessment of environmental noise - part 2. Second. 2007. Switzerland, International Standards Organization, (ISO).

(28) El Fadel M, Chahine M, Baaj H, Mezher T. Assessment of noise impacts at airports.

International Journal of Environmental Studies 2002; 59(4):447-467.

(29) Leong ST, Laortanakul P. Monitoring and assessment of daily exposure of roadside workers to traffic noise levels in an Asian city: a case study of Bangkok streets. Environmental Monitoring and Assessment 2003; 85:69-85.

(30) Miedema HME, Oudshoorn CGM. Annoyance from transportation noise: relationships with exposure metrics DNL and DENL and their confidence intervals. Environmental Health Perspectives 2001; 109(4):409-416.

(31) Gaja E, Gimenez A, Sancho S, Reig A. Sampling techniques for the estimation of the annual equivalent noise level under urban traffic conditions. Applied Acoustics 2003; 64:43-53.

(32) Ng CH, Tang SK. On monitoring community noise using arbitrarily chosen measurement periods. Applied Acoustics 2007.

(33) Environmental Protection Agency (EPA). Environmental Noise. 2007. http://epa.sa.gov.au/pdfs/info_noise.pdf (34) World Health Organization (WHO). Guidelines for Community Noise. 1999. http://www.who.int/docstore/peh/noise/guidelines2.html

(35) International Standards Organization. ISO 1996 - Acoustics - description, measurement and assessment of environmental noise - part 1. Second. 2003. Switzerland, International Standards Organization, (ISO).

(36) DMTI. Land use and City Roads, Halifax NS. Maps & Geospatial Information Collection,

Dalhousie University. Markham, Ontario. 1-9-2007. (37) Center Technology Corp. Instruction manual for Center 322 data logger sound level

meter. 2006. (38) Marantz. Marantz professional portable solid state recorder, model PMD660 user guide.

2005.

Page | 55

(39) McDonald JH. Handbook of Biological Statistics. University of Delaware . 5-1-2008. http://udel.edu/~mcdonald/statkruskalwallis.html (40) Heckert A. Levene Test. National Institute of Standards and Technology . 4-4-2003. http://www.itl.nist.gov/div898/software/dataplot/refman1/auxillar/levetest.htm

community environmental noise and the built environment in...

Documents